Voice Isolation and Dialogue Leveler Real Time Playback

When applying either of these AI effects to a track, they take a few seconds to kick in upon playback. This seems to be pretty common amongst other users.

Does anyone have real time playback with these effects applied to a track? If so, what kind of system are you running?

How do you manage your workflow with these effects? For example when making real time edits with a client, I have to toggle it off, but then need to export the timeline in order to hear the final mix without the delay?

Mac Studio Ultra has the same issue. I suspect what is happening is that when you make a cut, the audio effect needs to be recalculated for the result to make sense. This can take a while to run through the length of the clip, analyzing and recalculating each frame before caching it.

The easiest workflow is not to turn either on when the client is over your shoulder, and turn it on when you render a version.

I have two bodge methods 1) start the playback a few seconds before the point at which you wish to hear the target audio so that the AI has the chance to be ready; you could do this by making a temporary cut at e.g. 3 seconds before and then use the up arrow to start the playback from there 2) add VI at e.g. amount 50 and Dialogue Leveler at e.g. more lift for low levels to the entire section of audio track in question, export it as a WAV file, bring the new WAV file into the timeline beneath the original, cancel VI and DL on the original, then you can switch between an iteration of no VI+DL and applied VI+DL by Solo selecting each one in turn.

Messy for sure and you will probably want to tweak the VI+DL for your final version but at least you and your listener can appreciate the difference.

I can tell you that even the standalone iZotope RX Advanced has the same limitations. You basically need to give it time to render the results, so it's not a real-time process because it's too computationally intensive. I bet there's a way to cache the effect, but I'm not sure how that would work with just audio.

Voice Isolation may be cached when applied as a Clip effect.

Clip contextual menu > Cache Audio Effects